A mental model is a collection of internal factors such as motivation, thought processes, emotions, and needs. It serves as the standard by which users evaluate products. This study investigates the process and methods for constructing a mental model of user behavior using an eye tracker, designs an eye-tracking experiment for interface interaction behavior, and analyzes and organizes the data. Eye-tracking metrics are used to calculate users’ PAD emotional values and establish a user emotional prediction model. The relationship between eyetracking data and PAD multi-dimensional emotional values is explored, and the model’s validity is verified. The results show that the Sig. values of the emotional prediction model are all greater than 0.05, indicating high predictive capability and the ability to accurately predict users’ emotional preferences for web interfaces. Then, a user behavior HMM model, a user emotional PAD dimension model, and a user UI interface mapping model were established, and a mobile adaptive UI was designed for application verification
